OpenAI announced a neural network for voice cloning Voice Engine

OpenAI, known for its chatbot ChatGPT, is about to create its own voice assistant and has demonstrated technology for cloning a person's voice. At the moment, the company will not release it publicly due to security reasons.

Tech Mar 31, 2024 0 230 Add to Reading List

OpenAI announced a neural network for voice cloning Voice Engine

OpenAI unveiled its new Voice Engine technology on Friday, just over a week after filing a trademark application. The company claims it can recreate a person's voice using just 15 seconds of recording of their conversation.

Below are recordings from the test version of Voice Engine, the quality of which is almost indistinguishable from human speech.

OpenAI says it plans to conduct preliminary testing with early testers, but not release the technology into widespread use due to the risk of misuse.

We recognize that generating speech that sounds like people's voices carries serious risks, which are especially true in an election year
From an appeal from OpenAI

In New Hampshire, authorities are investigating robocalls made to thousands of voters shortly before the presidential primaries that used an artificial intelligence-generated voice that imitated US President Joe Biden's.

OpenAI previously proposed a tool that could create digital images based on description, and then created similar technology to create full-motion videos.

The company is already planning meetings with key Hollywood players to promote the new Sora video generator, and has announced a new version of ChatGPT will be released mid-year.